*4.2. Solution Preparation*

Milli-Q water (resistivity: 18.2 MΩ·cm; interfacial tension: 72.4 mN·m<sup>−</sup><sup>1</sup> at 22 ◦C; total organic carbon content: < 4 mg·L−1) was used to prepare all solutions and for cleaning of surfaces and glassware. The background electrolyte for the polyelectrolyte solutions was pH 5 0.1 M KCl solution (pH adjusted prior to making other solutions). PEI (500 ppm) was prepared in background electrolyte, stirred overnight and used within one week. Solutions of CS and FUC (both 500 ppm) were prepared in background electrolyte and stirred overnight. The polysaccharide solutions were used within 24 h of preparation. The background electrolyte, FUC and CS solutions were pH adjusted with volumetric grade KOH and HCl solutions to pH 5 before experiments. All pH adjustments were performed to give a value of ± 0.05 from the desired pH. PEI was used at its native pH in 0.1 M KCl pH 5 solution. Both LYZ and FGF-2 (both 25 ppm) were prepared the day of the experiment in PBS (pH 7.3) and stirred briefly to dissolve the protein.The concentration of 25 <sup>μ</sup>g·mL−<sup>1</sup> FGF-2 solution was chosen for two main reasons; (i) physiological concentrations are approximately 50 pg·mL−<sup>1</sup> in plasma [74], (ii) spectroscopic detection levels were found to be in the range of hundreds of <sup>μ</sup>g·mL−<sup>1</sup> for solution spectra. Typically, adsorption to a multilayer increases the concentration within the evanescent wave, and thus detection of less than this concentration is possible, it was decided to work in a range that would ensure detection. In addition other authors have found concentrations between 1.65–100 <sup>μ</sup>g·mL−<sup>1</sup> growth factor solutions su fficient for multilayer studies [8,18,75–77].

#### *4.3. Polyelectrolyte Multilayer Preparation and Growth Factor Adsorption*/*Incorporation*

Multilayers were prepared, in situ, under flow for all experiments. Initially, the system is flushed with KCl background electrolyte, then an anchoring layer of PEI is deposited by flowing the solution over the substrate for 15 min followed by a 5 min rinse with KCl. Following the PEI layer, FUC is adsorbed then CS. Each polymer is adsorbed for 15 min followed by a 5 min KCl rinse. The fucoidan and chitosan layers make one bilayer pair. This bilayer is repeated until the desired bilayer number is reached. Where FGF-2 was embedded in the film, a total of 8 bilayers were used with a FGF-2 layer at bilayer 6 i.e., PEI-(FUC/CS)5-(FUC/FGF-2)-(FUC/CS)2. For the permeation experiments multilayers composed of 9.5 bilayers were used so the film can be described by; PEI-(FUC/CS)9-FUC.

#### *4.4. Attenuated Total Reflectance Fourier Transform Infrared Spectroscopy (ATR FTIR)*

Fourier transform infrared experiments were performed on a Varian 670-IR FTIR spectrometer (Agilent Technologies, Mulgrave, VIC, Australia). A Ge internal reflection element (IRE) was used for FGF-2 permeation experiments. A ZnSe IRE was used for the data shown throughout the main manuscript. The IRE was mounted in a Fast IR single reflection ATR accessory (Harrick Scientific, Pleasantville, NY, USA) and fitted with a liquid flow cell attached to a peristaltic pump (Masterflex L/S, John Morris Scientific, Deepdene, VIC, Australia) with Tygon tubing (Masterflex L/S 13, Cole Parmer, Vernon Hills, IL, USA).

The ZnSe/Ge IRE (Harrick Scientific, Pleasantville, NY, USA) was bu ffed in a figure-of-eight pattern for approximately 5 min with OP-U colloidal silica suspension on a wet, MD-Nap ™ 250 mm polishing pad (both Struers, Ballerup, Denmark), followed by bu ffing for a further 2 min with Milli-Q water. Each component was sonicated in a surfactant for 30 min; the IRE in 2% pH 7 Tickopur, the flow cell and the tubing were sonicated in 2% Hellmanex - each solution was injected through the tubing 3 times with syringes (Luer slip, 5 cc·mL−1, Terumo, Tokyo, Japan). Each component was rinsed with Milli-Q water and then sonicated in 100% undenatured ethanol for 15 min (tubing was not exposed to ethanol), followed by a further rinse, then sonicated in Milli-Q water for 15 min. Finally, the components were rinsed a last time, dried under a stream of high purity dried nitrogen gas (99.999%, BOC, North Ryde, NSW, Australia) and allowed to dry fully overnight in a covered plastic container before being mounted.

Multilayers were created on the IRE surface under flow by following the protocol outlined above. The FGF-2 adsorption and PBS rinses were performed by flowing PBS over the multilayer for 5 min at 1.000 mL·min−1. Then the tubing was removed and FGF-2 solution was injected directly into the flowcell chamber via a syringe (Luer slip, 1 cc·mL−1, Terumo, Japan). The injection of the solution was staged over the 15 min adsorption, with 0.3 mL injected at 0, 5 and 10 min. For the di ffusion study the FGF-2 remained on the multilayer for 2 h. The tubing was reconnected to the flowcell and then flushed with PBS again for 5 min at 1.000 mL·min−1.

Single channel spectra from 256 scans were obtained in the region of 650 cm<sup>−</sup><sup>1</sup> (on the ZnSe IRE) or 780 cm<sup>−</sup><sup>1</sup> (on the Ge IRE) to 4000 cm<sup>−</sup>1, with 4 cm<sup>−</sup><sup>1</sup> resolution (commonly employed for studies of condensed matter systems, as higher resolution does not provide finer detail of peaks due to the natural linewidth of peaks in such systems) using Agilent Resolutions Pro software v5.2.0.36. Spectra were recorded for each experiment, as follows; (i) a background spectrum in air; (ii) a water vapour (WV) spectrum in air 10 min after the background spectrum; (iii) a spectrum of the background electrolyte after 5 min flow; (iv) then polymer spectra after each successive adsorption/rinse cycle.

Spectra were collected at specific time points, during the PBS/FGF-2/PBS cycle; (i) after the 5 min PBS rinse of the 5.5 bilayer, fucoidan terminating PEM, (ii) every 5 min during FGF-2 adsorption and (iii) after the 5 min PBS rinse after growth factor adsorption. These spectra were processed by subtracting the initial PBS rinse from the FGF-2 and subsequent PBS rinse with a 1 to 1 ratio, then a spectrum of PBS was added to flatten the O-H bending mode of water.

Whilst for the di ffusion study on the 9.5 bilayer multilayer, spectra were collected at 15 min FGF-2 adsorption and then after a final 5 min PBS rinse following the 2 h FGF-2 adsorption. These spectra were processed by subtracting the PBS rinse of the multilayer in a 1 to 1 ratio from all subsequent spectra. Each experiment was performed as two independent repeats. Spectral processing was performed with OMNIC software v8.2.0.387 (Thermo Fisher Scientific, Scoresby, VIC, Australia).

The multilayer build-up spectra presented below (Figure 3) and in the electronic supplementary material were produced by subtracting the spectrum of the background electrolyte (KCl) from each spectrum to remove the contribution of water in the O-H bending mode region (~1630 cm<sup>−</sup>1). In Figures 4 and 5 the spectra were produced by subtracting the spectra of the PBS rinse of the 6th fucoidan layer (PBS rinse after FUC6) in a 1 to 1 ratio from each, followed by adding a PBS spectra to flatten the region between 1650–1700 cm<sup>−</sup><sup>1</sup> to remove the O-H bending mode of water. Finally, manual water vapour correction was performed, followed by automatic baseline correction, for all spectra.

#### *4.5. Quartz Crystal Microbalance with Dissipation Monitoring (QCM-D)*

The experiments were performed on a Q-sense E4 instrument (Biolin Scientific, Västra Frölunda, Sweden) under continuous flow conditions. The multilayers are formed on Si-coated 5 MHz AT-cut quartz crystal sensors (SiO2 50 nm, QSX 303, Q-sense, Biolin Scientific, Sweden). The sensors were cleaned by sonicating in 1 M HCl for 30 min, followed by 2% Hellmanex for 30 min then Milli-Q water for 10 min. The sensors were individually dried under a stream of high purity dried nitrogen and air plasma cleaned for 60 s (Harrick Plasma, Ithaca, NY, USA).

Once cleaned, the sensors were placed into the QCM chambers, where they were allowed to stabilise in background electrolyte prior to measurement for 1 h under flow at 0.050 mL·min−<sup>1</sup> using a multi-channel peristaltic pump (Ismatec, Cole-Palmer, Wertheim, Germany). Then solutions were pumped through the system following the adsorption protocol described above at rates of 0.100 mL·min-<sup>1</sup> for polyelectrolyte/protein adsorption and 0.300 mL·min−<sup>1</sup> for the background electrolyte rinse.

The PBS/FGF-2/PBS cycle was performed using a flow rate of 0.300 mL·min−<sup>1</sup> for 5 min for the PBS rinse prior to the adsorption of growth factor. The FGF-2 adsorption was performed by flowing the growth factor solution over the multilayer for 1 min at a rate of 0.30 mL·min−1, then for 14 min at a rate of 0.05 mL·min−1. The subsequent PBS flush was performed at 0.05 mL·min−<sup>1</sup> and 0.30 mL·min−<sup>1</sup> for 5 min each (the extra slow flush was used to account for the additional exposure time of FGF-2 due to spectra collection in ATR FTIR spectroscopy measurements).
