Site-Specific Tryptophan Labels Reveal Local Microsecond

molecules

Article Site-Speciﬁc Tryptophan Labels Reveal Local Microsecond–Millisecond Motions of Dihydrofolate Reductase

Morgan B. Vaughn 1,*, Chloe Biren 2, Qun Li 2, Ashwin Ragupathi 2 and R. Brian Dyer 2,*

1 Division of Natural Sciences, Oglethorpe University, Atlanta, GA 30319, USA 2 Department of Chemistry, Emory University, Atlanta, GA 30322, USA; [email protected] (C.B.); [email protected] (Q.L.); [email protected] (A.R.) * Correspondence: [email protected] (M.B.V.); [email protected] (R.B.D.); Tel.: +1-404-504-1214 (M.B.V.); +1-404-727-6637 (R.B.D.) Academic Editor: Raﬁk Karaman Received: 24 July 2020; Accepted: 21 August 2020; Published: 22 August 2020

Abstract: Many enzymes are known to change conformations during their catalytic cycle, but the role of these protein motions is not well understood. Escherichia coli dihydrofolate reductase (DHFR) is a small, flexible enzyme that is often used as a model system for understanding enzyme dynamics. Recently, native tryptophan fluorescence was used as a probe to study micro- to millisecond dynamics of DHFR. Yet, because DHFR has five native tryptophans, the origin of the observed conformational changes could not be assigned to a specific region within the enzyme. Here, we use DHFR mutants, each with a single tryptophan as a probe for temperature jump fluorescence spectroscopy, to further inform our understanding of DHFR dynamics. The equilibrium tryptophan fluorescence of the mutants shows that each tryptophan is in a different environment and that wild-type DHFR fluorescence is not a simple summation of all the individual tryptophan fluorescence signatures due to tryptophan–tryptophan interactions. Additionally, each mutant exhibits a two-phase relaxation profile corresponding to ligand association/dissociation convolved with associated conformational changes and a slow conformational change that is independent of ligand association and dissociation, similar to the wild-type enzyme. However, the relaxation rate of the slow phase depends on the location of the tryptophan within the enzyme, supporting the conclusion that the individual tryptophan fluorescence dynamics do not originate from a single collective motion, but instead report on local motions throughout the enzyme.

Keywords: enzyme; dynamic; conformation; motion; relaxation; spectroscopy; ﬂuorescence; tryptophan; DHFR

1. Introduction Enzyme dynamics across a wide range of timescales are important for enzymatic catalysis. Conformational changes and protein motions have been implicated in every step of catalysis from crossing the barrier of the chemistry step on the femtosecond timescale to rotating whole domains on the order of milliseconds [1]. Dihydrofolate reductase (DHFR) is a classic model for studying enzyme dynamics and remains an active area of research [2–9]. DHFR is a small flexible enzyme with multiple mobile loops. In particular, the Met20 loop changes conformation throughout the catalytic cycle: it is closed in the substrate-bound complexes and occluded in the product-bound complexes [10]. Enzyme dynamics are paramount for DHFR function. It is well established that mutations that decrease loop flexibility impair DHFR enzymatic activity [11,12], but important conformational changes are not limited to the flexible loops. Molecular dynamics simulations and bioinformatics have proposed a network of correlated motions that extends throughout DHFR [13]. Experimental evidence supports

Molecules 2020, 25, 3819; doi:10.3390/molecules25173819 www.mdpi.com/journal/molecules Molecules 2020, 25, x FOR PEER REVIEW 2 of 13 Molecules 2020, 25, 3819 2 of 13 network of correlated motions that extends throughout DHFR [13]. Experimental evidence supports thatthat G121, G121, F125, F125, M42, M42, and and I14 I14 are are part part of of the the network network of of correlated correlated amino amino acids, acids, demonstrating demonstrating that that motionsmotions near near and and far far from from the the active active site site are are important important for for catalysis catalysis [ 6[6].]. Recently,Recently, temperature temperature jump jump (T-jump) (T-jump) fluorescence fluorescence spectroscopy spectroscopy has has been been used used to to examine examine the the dynamicsdynamics of of DHFR DHFR on theon microsecond the microsecond to millisecond to mill timescaleisecond [timescale14]. Laser-induced [14]. Laser-induced T-jump spectroscopy T-jump isspectroscopy unique in that is it canunique access in dynamics that it can from access tens of nanosecondsdynamics from to several tens of milliseconds. nanoseconds This to complete several timemilliseconds. regime is difficultThis complete to access time with regime techniques is difficult traditionally to access used with to techniques study enzyme traditionally dynamics used such to asstudy NMR enzyme spectroscopy, dynamics hydrogen such as deuteriumNMR spectroscopy exchange, hydrogen (HDX), stopped-flow, deuterium exchange and molecular (HDX), dynamics stopped- simulationsflow, and molecular [15–18]. Additionally, dynamics simulations because T-jump [15–18]. is aAdditionally, type of relaxation because methodology, T-jump is analyzinga type of T-jumprelaxation transients methodology, provides analyzin informationg T-jump about transients both the forwardprovides and information reverse processes, about both which the forward allows usand to reverse determine processes, the relationship which allows between us to relaxation determine events the relationship based on theirbetween concentration relaxation dependence.events based Previously,on their concentration we measured dependence. the dynamics Previously, of wt-DHFR we withmeasured T-jump the spectroscopy dynamics of usingwt-DHFR tryptophan with T-jump (Trp) fluorescencespectroscopy as ourusing probe tryptophan [14]. DHFR (Trp) has fivefluorescen native tryptophansce as our (W22,probe W30,[14]. W47,DHFR W74, has W133) five located native throughouttryptophans the (W22, enzyme, W30, as W47, shown W74, in Figure W133)1. W22located is located throughout on the the Met20 enzyme, loop; W133as shown is located in Figure at the 1. terminusW22 is located of a beta on sheetthe Met20 near theloop; FG W133 loop; is W30 located is near at the substrate-bindingterminus of a beta site;sheet and near W47 the and FG W74loop; areW30 close is near together the substrate-binding but removed from site; the and active W47 site. and We W74 observed are close relaxation together rates but onremoved a microsecond from the timescaleactive site. strongly We observed coupled torelaxation ligand association rates on /adissociation microsecond and timescale a slower conformationalstrongly coupled change to ligand that wasassociation/dissociation not correlated to ligand and binding. a slower Since conformational wild-type DHFR change (wt-DHFR) that was has not multiple correlated tryptophans, to ligand thebinding. previous Since study wild-type could not DHFR identify (wt-DHFR) the origin has of mult the observediple tryptophans, slow conformational the previous change. study could not identify the origin of the observed slow conformational change.

FigureFigure 1. 1.Crystal Crystal structurestructure (Protein(Protein DataData Bank: Bank: 1RX2)1RX2) ofof wild-typewild-type E.E. coli coli dihydrofolate dihydrofolate reductase reductase (DHFR)(DHFR) (magenta) (magenta) with with oxidized oxidized nicotinamide nicotinamide adenine adenine dinucleotide dinucleotide phosphate phosphate (NADP (NADP+)+) cofactor cofactor (blue)(blue) and and folate folate (yellow). (yellow). The The ﬁve five native native tryptophan tryptophan (Trp) (Trp) residues residues are are labeled labeled and and shown shown in in green. green.

Here,Here, wewe characterizecharacterize thethe equilibriumequilibrium ﬂuorescencefluorescence andand T-jumpT-jump dynamicsdynamics ofof DHFRDHFR mutants,mutants, eacheach with with a a single single tryptophan tryptophan residue residue (denoted (denoted asasmidW midWmutants). mutants). WeWe foundfound thatthat allallof ofthe the midW midW mutantsmutants exhibited exhibited biphasic biphasic relaxation relaxation rates rates similar similar to to wt-DHFR wt-DHFR with with a a concentration-dependent concentration-dependent fast fast relaxationrelaxation rate rate and and a concentration-independenta concentration-independent slow slow relaxation relaxation rate. rate. Notably, Notably, the slowthe slow relaxation relaxation rate ofrate the of midW the midW mutants mutants diﬀers differs based onbased the locationon the loca of thetion Trp of probethe Trp within probe the within enzyme the and enzyme is generally and is fastergenerally than faster the slow than event the slow observed event previously observed inprev wt-DHFR.iously in This wt-DHFR. evidence This supports evidence the supports conclusion the thatconclusion the observed that the relaxation observed rate relaxation in wt-DHFR rate likely in wt-DHFR arises from likely conformational arises from conformational changes that modulate changes thethat Trp–Trp modulate interactions the Trp–Trp and thatinteractions individual and tryptophan that individual residues tryptophan report on local residues motions. report on local motions.

2. Results and Discussion

2.1. Protein Characterization

Molecules 2020, 25, x FOR PEER REVIEW 3 of 13

To verify that the Trp to phenylalanine (Phe) mutations did not adversely affect the stability and function Moleculesof midW2020 ,mutant25, 3819 DHFR enzymes, the melting temperature and activity of the enzymes3 of 13 were examined. The effects of the Trp to Phe mutations on enzyme catalysis were determined by comparing2. Resultswt-DHFR and Discussionand midW mutant activity by monitoring nicotinamide adenine dinucleotide phosphate2.1. (NADPH) Protein Characterization absorbance at 340 nm. As shown in Table S1, the midW22 mutant retained 90% activity compared to the wild-type enzyme. The high activity of midW22 compared to wt-DHFR To verify that the Trp to phenylalanine (Phe) mutations did not adversely affect the stability and demonstrates that W30, W47, W74, and W133 have little importance in the catalytic activity of DHFR, function of midW mutant DHFR enzymes, the melting temperature and activity of the enzymes were since theexamined. midW22 Themutant effects has of the all Trp of tothe Phe Trp mutations residues, on enzymeexcept catalysisW22 is werereplaced determined with byPhe. comparing The other four midW mutantswt-DHFR have and midWreduced mutant activi activityty (20–33%), by monitoring which nicotinamide can be explai adeninened dinucleotideby disrupting phosphate the hydrogen bond network(NADPH) between absorbance W22, at D27, 340 nm. T113, As shownand the in substrate Table S1, the DHF midW22 [19]. mutant retained 90% activity Circularcompared dichroism to the wild-type (CD) spectroscopy enzyme. The high was activity used of midW22to examine compared the tofold wt-DHFR and demonstratesthe stability of the that W30, W47, W74, and W133 have little importance in the catalytic activity of DHFR, since the midW mutants. Figure 2A shows the CD spectra. Notably, the molar ellipticity of the midW mutants midW22 mutant has all of the Trp residues, except W22 is replaced with Phe. The other four midW is lower comparedmutants have to reduced the wild-type activity (20–33%), enzyme. which In the can midW be explained mutants, by disrupting all but one the of hydrogen the five bond native Trp residues networkhave been between mutated W22, D27,to Phe. T113, Previously, and the substrate sing DHFle Trp [19 to]. Phe, valine, or leucine mutations have been shown Circularto impact dichroism the CD (CD) spectra spectroscopy from was 190 used to to 250 examine nm the[20]. fold The and themutations stability of caused the midW a loss of ellipticity,mutants. gain of Figure ellipticity,2A shows or the a mixture CD spectra. of Notably,the two the in molardifferent ellipticity regions of the of midW the CD mutants spectra is lower depending compared to the wild-type enzyme. In the midW mutants, all but one of the five native Trp residues on the location of the Trp mutation and which amino acid was chosen to replace the Trp residue. In have been mutated to Phe. Previously, single Trp to Phe, valine, or leucine mutations have been shown particular,to impactexciton the coupling CD spectra between from 190 W47 to 250 and nm [W7420]. The has mutations a large causedimpact a on loss the of ellipticity, CD spectra gain of DHFR. In each ofellipticity, the midW or a mutants, mixture of W47, the two W74, in diff orerent both regions have of been the CD removed, spectra depending making ona direct the location comparison of to the wt-DHFRthe Trp spectra mutation difficult and which to amino interpret. acid was However, chosen to replace the CD the Trpspectra residue. does In particular, show that exciton the midW mutants couplinghave a minimum between W47 band and W74in molar has a largeellipticity impact around on the CD 215 spectra nm ofand DHFR. a maximum In each of thearound midW 195 nm, mutants, W47, W74, or both have been removed, making a direct comparison to the wt-DHFR spectra which is characteristic for folded proteins with both alpha helices and beta sheets. The melting difficult to interpret. However, the CD spectra does show that the midW mutants have a minimum temperaturesband inof molarthe midW ellipticity mutants around were 215 nm determined and a maximum by monitoring around 195 nm,the whichloss of is ellipticity characteristic at 222 nm with increasingfor folded proteinstemperatures with both via alpha CD helices spectros and betacopy sheets. (Figure The melting 2B). temperaturesWe determined of the midWthe melting temperaturemutants of werewt-DHFR determined to be by 48 monitoring ± 1 °C, the which loss of is ellipticity similar at to 222 the nm withpreviously increasing determined temperatures melting via CD spectroscopy (Figure2B). We determined the melting temperature of wt-DHFR to be 48 1 C, temperature of 49.3 °C [21]. The melting temperatures of the midW mutants range from± 42◦ to 49 °C which is similar to the previously determined melting temperature of 49.3 C[21]. The melting (Table 1). The melting temperature is an important consideration, since the◦ proteins are transiently temperatures of the midW mutants range from 42 to 49 ◦C (Table1). The melting temperature is heated to 36 °C during the T-jump experiments. For midW22, midW47, midW74, and midW133, the an important consideration, since the proteins are transiently heated to 36 ◦C during the T-jump final temperatureexperiments. is 9 For °C midW22, or more midW47, below midW74,the melting and midW133,temperature, the final so temperatureunfolding is is 9not◦C likely or more to play a large role.below MidW30 the melting has a temperature, lower melting so unfolding temperature is not likelyof 42 to°C, play so aunfolding large role. MidW30effects may has a be lower relevant to the observedmelting dynamics. temperature of 42 ◦C, so unfolding effects may be relevant to the observed dynamics.

Figure 2. (A) Circular dichroism (CD) spectra of wild-type DHFR (wt-DHFR) and the ﬁve midW Figure 2.mutants. (A) Circular Trp side dichroism chains can impact (CD) the spectra CD spectra. of wild Wt-DHFR-type has DHFR ﬁve trp (wt-DHFR) residues, whereas and the the midW five midW mutants.mutants Trp side have chains only one can trp impact each. Despite the CD the spec lowertra. molar Wt-DHFR ellipticity has of the five mutants trp residues, compared towhereas the the midW mutantswild type, have the CDonly spectra one showtrp each. that the Despite mutants the are lower folded enzymesmolar ellipticity with both alpha of the helices mutants and beta compared to the wildsheets. type, (B )the Thermal CD spectra melts of wt-DHFRshow that and the the mu midWtants mutants. are folded midW: enzymes DHFR mutants, with both each withalpha a helices single tryptophan residue. and beta sheets. (B) Thermal melts of wt-DHFR and the midW mutants. midW: DHFR mutants, each with a single tryptophan residue.

Table 1. Melting temperatures of wild-type (WT) and midW mutants as determined by loss of ellipticity at 222 nm.

Enzyme midW22 midW30 midW47 midW74 midW133 WT Melting Temperature (°C) 46 ± 2 42 ± 1 49 ± 2 45 ± 2 45 ± 2 48 ± 1

Molecules 2020, 25, x FOR PEER REVIEW 4 of 13

2.2. Equilibrium Fluorescence The tryptophan fluorescence spectra of all five midW mutants were measured for the apoenzyme and three different enzyme·ligand complexes: E·Folate, E·NADP+, and E·NADP+·Folate. Each of these enzyme–ligand complexes are model systems for the different enzyme states in the Molecules 2020, 25, 3819 4 of 13 catalytic cycle: the binary product complex, the holoenzyme, and the Michaelis complex, respectively. The fluorescence intensity of each midW mutant was much less than the wild-type enzyme (Figure S2), Tableas expected, 1. Melting since temperatures the midW of wild-type mutants (WT) each and contain midW mutantsonly one as determinedtryptophan, by losswhereas of ellipticity wt-DHFR containsat 222 5 nm.tryptophans. As shown in Figure 3A, the intensity of the Trp fluorescence of each midW mutant differs, with midW74 exhibiting the greatest fluorescence intensity, followed by midW30, Enzyme midW22 midW30 midW47 midW74 midW133 WT midW47, midW133, and midW22 with the lowest fluorescence intensity. Additionally, the shift in the Melting Temperature ( C) 46 2 42 1 49 2 45 2 45 2 48 1 Trp fluorescence maximum◦ between± midW mutants± sh±ows that the± Trp in each± mutant± is in a different environment. Typically, a blue shift indicates a more hydrophobic environment and tends 2.2.to be Equilibrium concomitant Fluorescence with an increase in fluorescence intensity. A red shift indicates a more hydrophilic environment,The tryptophan which fluorescenceusually correspon spectrads of with all fivea decrease midW mutantsin fluorescence were measured intensity, for due the to apoenzyme quenching andby water. three diInterestingly,fferent enzyme theligand Trp fluorescence complexes: of E Folate, the midW E NADP mutants+, and does E NADP not necessarily+ Folate. Each follow of thesethese · · · · · enzyme–ligandtrends. For example, complexes midW47 are and model midW133 systems are for blue the shifted different compared enzyme to states midW74 in the and catalytic midW30, cycle: yet themidW47 binary and product midW133 complex, display the holoenzyme,lower fluorescence and the intens Michaelisities.complex, This indicates respectively. that there The must fluorescence be some intensityquenching of eachinteractions midW mutantwith W47 was and much W133 less thanin the the apoenzyme. wild-type enzyme Additionally, (Figure S2),there as are expected, likely sincequenching the midW interactions mutants eachat position contain 22, only given one tryptophan, that midW22 whereas exhibits wt-DHFR a similar contains shift 5compared tryptophans. to AsmidW74 shown and in Figure midW30.3A, the The intensity similar of shift the Trp suggests fluorescence a similar of each environment; midW mutant however, di ffers, withmidW22 midW74 Trp exhibitingfluorescence the is greatest drastically fluorescence reduced. intensity, followed by midW30, midW47, midW133, and midW22 with theFigure lowest 3B fluorescenceshows the comparison intensity. Additionally, of the wt-DHFR the shift Trp in fluorescence the Trp fluorescence spectrum maximum to the sum between of the midWTrp fluorescence mutants shows spectra that of the the Trp midW in each mutants. mutant The is in intensities a different have environment. been normalized Typically, to a 1 blue to better shift indicatesdemonstrate a more the hydrophobic difference in environment lambda max and between tends to the be concomitanttwo spectra. with The an difference increase inin fluorescence the spectra intensity.shows that A the red wt-DHFR shift indicates Trp fluorescence a more hydrophilic is not a environment,simple linear whichcombination usually of corresponds the fluorescence with of a decreasethe individual in fluorescence tryptophans. intensity, This due has to been quenching observed by water.previously, Interestingly, albeit in the a Trpless fluorescencedirect manner of theby midWOhmae mutants et al., where does not a single necessarily tryptophan follow thesewas muta trends.ted Forto various example, hydrophobic midW47 and residues midW133 [20]. are They blue shiftedobserved compared that the toTrp midW74 fluorescence and midW30, decreased yet in midW47 varying andamounts midW133 depending display on lower which fluorescence tryptophan intensities.was mutated. This In indicates this way, that they there came must to the be same some conclusion quenching as interactions stated above. with Furthermore, W47 and W133 W47 in and the apoenzyme.W74 are known Additionally, to form an there excimer are likely pair [20]. quenching Typically, interactions the emission at position of the 22, excimer given thatis red midW22 shifted exhibitscompared a similarto the emission shift compared of the fluorescent to midW74 mono andmers, midW30. which The may similar contribute shift to suggests the observed a similar red environment;shift of the wild-type however, DHFR midW22 fluorescence Trp fluorescence compared is drastically to the sum reduced. of the fluorescence of the individual midW mutants.

FigureFigure 3.3. ((A)) TrpTrp fluorescencefluorescence ofof thethe apoenzymesapoenzymes ofof thethe fivefive midWmidW mutantsmutants (3(3 µμM).M). DiDifferencesfferences inin fluorescencefluorescence intensity intensity and and shift shift indicate indicate that that each ea tryptophanch tryptophan is in is a uniquein a unique environment. environment. (B) Wt-DHFR (B) Wt-

TrpDHFR fluorescence Trp fluorescence spectrum spectrum and the and summation the summation of the midW of the mutantsmidW mutants0 fluorescence′ fluorescence spectra. spectra. Wt-DHFR Wt- hasDHFR a much has greatera much fluorescencegreater fluorescence intensity intensity than the individualthan the individual midW mutants midW combined. mutants combined. The intensities The haveintensities been normalizedhave been normalized here to better here demonstrate to better demonstrate the blue shift the of theblue summed shift of spectrathe summed compared spectra to thecompared wild-type to the spectrum. wild-type This spectrum. data show This that data the sh wt-DHFRow that the Trp wt-DHFR fluorescence Trp isfluorescence not a simple is linearnot a combinationsimple linear of combination each individual of each tryptophan’s individual fluorescence.tryptophan’s fluorescence.

Figure3B shows the comparison of the wt-DHFR Trp ﬂuorescence spectrum to the sum of the

Trp fluorescence spectra of the midW mutants. The intensities have been normalized to 1 to better demonstrate the difference in lambda max between the two spectra. The difference in the spectra shows that the wt-DHFR Trp fluorescence is not a simple linear combination of the fluorescence of the individual tryptophans. This has been observed previously, albeit in a less direct manner Molecules 2020, 25, x FOR PEER REVIEW 5 of 13

MoleculesTwo2020 patterns, 25, 3819 emerge when the Trp fluorescence of the midW mutant ligand complexes5 ofare 13 compared. Figure 4A shows the midW22 mutant complexes, which exhibit the same trend as all the other mutants except for midW47, which is shown in panel B. Table 2 shows the integrated by Ohmae et al., where a single tryptophan was mutated to various hydrophobic residues [20]. fluorescence intensity of each complex as a percentage of the apoenzyme fluorescence intensity. They observed that the Trp fluorescence decreased in varying amounts depending on which tryptophan MidW22 and the other three midW mutants (midW30, midW74, and midW133) show very little was mutated. In this way, they came to the same conclusion as stated above. Furthermore, W47 and change upon binding NADP+ compared to the apoenzyme and show a significant change compared W74 are known to form an excimer pair [20]. Typically, the emission of the excimer is red shifted to apoenzyme for the binary complex with folate and the tertiary complex. Importantly, there is only compared to the emission of the fluorescent monomers, which may contribute to the observed red a small difference in the Trp fluorescence intensity between the E·Folate and E·NADP+·Folate shift of the wild-type DHFR fluorescence compared to the sum of the fluorescence of the individual complexes, which exist in the occluded and closed conformations, respectively. This indicates that midW mutants. W22, W30, W74, and W133 are not particularly sensitive to the closed versus occluded conformational Two patterns emerge when the Trp fluorescence of the midW mutant ligand complexes are states of the Met20 loop and are instead primarily sensitive to the presence of folate. Conversely, the compared. Figure4A shows the midW22 mutant complexes, which exhibit the same trend as all the Trp fluorescence of midW47 (Figure 4B) decreases by different amounts based on which ligands are other mutants except for midW47, which is shown in panel B. Table2 shows the integrated fluorescence bound. There is a small decrease upon binding NADP+, a larger decrease upon binding folate, and a intensity of each complex as a percentage of the apoenzyme fluorescence intensity. MidW22 and the further decrease in the tertiary complex. However, the combined decrease in fluorescence intensity other three midW mutants (midW30, midW74, and midW133) show very little change upon binding of individually binding folate and NADP+ is less than the total decrease observed for the tertiary NADP+ compared to the apoenzyme and show a significant change compared to apoenzyme for the complex. This means that the decrease in fluorescence intensity observed in the tertiary complex is binary complex with folate and the tertiary complex. Importantly, there is only a small difference in not simply due to quenching of the Trp fluorescence from the presence of ligands; there is also a the Trp fluorescence intensity between the E Folate and E NADP+ Folate complexes, which exist in the contribution from the conformation of the Met20· loop. Thus,· midW47· is more sensitive to changes in occluded and closed conformations, respectively. This indicates that W22, W30, W74, and W133 are the enzyme conformation compared to the other midW mutants. not particularly sensitive to the closed versus occluded conformational states of the Met20 loop and are instead primarily sensitive to the presence of folate. Conversely, the Trp fluorescence of midW47 Table 2. Integrated Trp fluorescence intensity of the enzyme–ligand complexes as a percentage of the (Figure4B) decreases by di fferent amounts based on which ligands are bound. There is a small decrease integrated Trp fluorescence intensity of the apoenzyme. upon binding NADP+, a larger decrease upon binding folate, and a further decrease in the tertiary complex. However, the combined decreaseE·NADP in fluorescence+ E·Folate intensityE·NADP+ of·Folate individually binding folate and NADP+ is less than the totalWild-type decrease observed93% for 36%the tertiary complex. 24% This means that the decrease in fluorescence intensity observedmidW22 in the102% tertiary complex 52% is not 49% simply due to quenching of the Trp midW30 106% 42% 40% fluorescence from the presence of ligands; there is also a contribution from the conformation of the midW47 92% 71% 54% Met20 loop. Thus, midW47midW74 is more sensitive98% to changes 64% in the enzyme 61% conformation compared to the other midW mutants. midW133 93% 64% 63%

FigureFigure 4. 4. EquilibriumEquilibrium Trp Trp fluorescence fluorescence of of midW22 midW22 ( (AA)) and and midW47 midW47 ( (BB)) and and their their ligand ligand complexes complexes (3(3 μµMM enzyme enzyme with with 6 6 μµMM folate folate and/or and/or 6 μµM NADP +).). The The Trp Trp fluorescence fluorescence of of midW22 midW22 appears appears to to be be primarilyprimarily modulated modulated by by the the presence presence of of folate, folate, whereas whereas the the Trp Trp fluorescence fluorescence of of midW47 midW47 changes changes with with eacheach ligand ligand complex. complex.

TheThe temperature-dependent Trp Trp fluorescence fluorescence of theof the midW midW mutants mutants is shown is shown in Figure in Figure5. The values5. The valueshave been have corrected been corrected for the intrinsicfor the intrinsic temperature temperature dependence dependence of Trp fluorescence, of Trp fluorescence, which decreases which decreaseswith increasing with increasing temperature temperature (Figure S3). (Figure This S3). is opposite This is opposite of the trend of the observed trend observed for the enzymefor the enzymecomplexes. complexes. All of All the of midW the midW mutants mutants follow follow the the same same general general trend trend as as thethe wild-type enzyme: enzyme: an increase in fluorescence intensity with increasing temperature, due primarily to ligand dissociation

Molecules 2020, 25, 3819 6 of 13 and accompanying loop conformation changes. Each of the midW mutants follow the trend as expected based on their change in Trp fluorescence with different ligands bound. An additional contribution to the increased Trp fluorescence is likely caused by loosening of the enzyme structure at elevated temperatures, which in turn reduces quenching interactions within the protein. This explains the increase in Trp fluorescence for the apoenzyme where no ligands are present to dissociate. Figure5 shows the change in fluorescence intensity for midW22 and midW47 as representative data.

MoleculesTable 2020 2., 25Integrated, x FOR PEER Trp REVIEW ﬂuorescence intensity of the enzyme–ligand complexes as a percentage of the6 of 13 integrated Trp ﬂuorescence intensity of the apoenzyme. an increase in fluorescence intensity with increasing temperature, due primarily to ligand E NADP+ E Folate E NADP+ Folate dissociation and accompanying loop ·conformation· changes. ·Each of· the midW mutants follow the trend as expected basedWild-type on their change 93% in Trp 36% fluorescence with 24% different ligands bound. An additional contribution tomidW22 the increased 102% Trp fluorescence 52% is likely caused 49% by loosening of the enzyme midW30 106% 42% 40% structure at elevated temperatures,midW47 which 92% in turn re 71%duces quenching 54% interactions within the protein. This explains the increasemidW74 in Trp fluorescence 98% for 64%the apoenzyme 61% where no ligands are present to dissociate. Figure 5 showsmidW133 the change 93% in fluorescence 64% intensity 63% for midW22 and midW47 as representative data.

Figure 5. ChangeChange in in fluorescence fluorescence intensity intensity versus versus temper temperature,ature, corrected corrected for for the the intrinsic intrinsic temperature temperature dependence ofof TrpTrp fluorescence, fluorescence, of of (A ()A midW22) midW22 and and (B) midW47(B) midW47 (3 µ M(3 enzymeμM enzyme with 6withµM folate6 μM andfolate/or and/or6 µM NADP 6 μM+ ).NADP All complexes+). All complexes demonstrate demonstrate an increase an in increa fluorescencese in fluorescence with increasing with temperature. increasing temperature.The change in The intensity change for in intensit the binaryy for complex the binary with complex NADP +withand NADP apoenzyme+ and apoenzyme compared tocompared the two tofolate the two bound folate complexes bound complexes is similar is for similar midW22, for mi whichdW22, is which expected is expected based on based the on similarities the similarities in the inspectra the spectra of apoenzyme of apoenzyme/E·NADP/E NADP+ and E +Folate and /E·Folate/E·NADPE NADP+ Folate.+·Folate. Conversely, Conversely, each midW47 each midW47 complex · · · · complexhas a slightly has a dislightlyfferent different temperature temperature dependence, dependen reflectingce, reflecting the difference the difference in Trp fluorescence in Trp fluorescence intensity intensityfor the di forfferent the liganddifferent bound ligand states. bound states.

2.3. Temperature Temperature Jump Enzyme Dynamics The T-jump T-jump transients transients for for all all five ﬁve midW midW mutants mutants with with folate folate share share a few a fewkey characteristics. key characteristics. The transientsThe transients fit to ﬁt double to double exponentials exponentials with with two twodistinct distinct phases phases (Table (Tables S3 and S3 andS4). S4).To determine To determine the originthe origin of the of observed the observed relaxation relaxation rates, rates, we can we canexamine examine the correlation the correlation between between the sum the sumof the of concentrationsthe concentrations of the of free the enzyme free enzyme and free and ligand. free ligand. Since T-jump Since T-jumpspectroscopy spectroscopy is a relaxation is a relaxation method thatmethod provides that providesinformation information about both about theboth forward the forwardand reverse and reversereactions, reactions, if the observed if the observed rate is concentrationrate is concentration dependent, dependent, then there then must there be must a bimo belecular a bimolecular step—in step—in this case, this ligand case, binding. ligand binding. Figure 6Figure shows6 showsrepresentative representative T-jump T-jump transients transients of midW of midW4747 and the and concentration-dependent the concentration-dependent plots. plots. The sumThe sumof free of freeconcentrations concentrations at 36 at °C, 36 the◦C, temperat the temperatureure after after the laser the laser pulse, pulse, was was calculated calculated using using the theKd andKd and ∆H ∆ofH folate of folate binding binding determined determined by isothermal by isothermal titration titration calorimetry calorimetry (ITC). (ITC). The TheITC ITCresults results are summarized in Table S2. The concentration dependence of the fast and slow relaxation rates for the midW mutants are shown in Figure 6 and Figure S4. The fast relaxation rate of the midW mutants are positively correlated with the sum of the concentrations of the free enzyme and free ligand. The strength of the correlation can be quantitatively described by the linear correlation coefficient, r. The r values range from 0.867 to 0.954 for midW30, midW47, midW74, and midW133. The threshold for significance at a 99% confidence level is 0.505 (DF = 23). [22] Thus, all of the midW mutants except for midW22 are significantly correlated with concentration and can be qualitatively described as strongly concentration dependent. MidW22 exhibits a weak concentration dependence with an r value of 0.366, which is higher than the cutoff for significance at the 90% confidence level (0.337, degrees of freedom (DF) = 23). The concentration dependence of the fast relaxation rate indicates that

Molecules 2020, 25, 3819 7 of 13 are summarized in Table S2. The concentration dependence of the fast and slow relaxation rates for the midW mutants are shown in Figure6 and Figure S4. The fast relaxation rate of the midW mutants are positively correlated with the sum of the concentrations of the free enzyme and free ligand. The strength of the correlation can be quantitatively described by the linear correlation coefficient, r. The r values range from 0.867 to 0.954 for midW30, midW47, midW74, and midW133. The threshold for significance at a 99% confidence level is 0.505 (DF = 23). [22] Thus, all of the midW mutants except for midW22 are significantly correlated with concentration and can be qualitatively described as strongly concentration dependent. MidW22 exhibits a weak concentration dependence with an r value of 0.366, which is higher than the cutoff for significance at the 90% confidence level (0.337, degrees of freedom (DF)Molecules= 23). 2020 The, 25 concentration, x FOR PEER REVIEW dependence of the fast relaxation rate indicates that the fast relaxation7 of 13 phase is convolved with ligand binding/unbinding. Conversely, the correlation coefficients of the the fast relaxation phase is convolved with ligand binding/unbinding. Conversely, the correlation slow relaxation rate vary from 0.659 to 0.124. A negative correlation with regard to concentration is coefficients of the slow relaxation− rate vary from −0.659 to 0.124. A negative correlation with regard nonsensical. Therefore, we can conclude that the slow relaxation rate of midW22, midW30, midW47, to concentration is nonsensical. Therefore, we can conclude that the slow relaxation rate of midW22, and midW133 is not concentration dependent. MidW74 has an r value of 0.124, which is a very low midW30, midW47, and midW133 is not concentration dependent. MidW74 has an r value of 0.124, correlation coefficient. Even at an 80% confidence level, the midW74 correlation coefficient is lower which is a very low correlation coefficient. Even at an 80% confidence level, the midW74 correlation than the threshold value (0.265, DF = 23), indicating that there is no significant correlation between coefficient is lower than the threshold value (0.265, DF = 23), indicating that there is no significant the slow relaxation rate and the sum of the free concentrations. The concentration dependence of correlation between the slow relaxation rate and the sum of the free concentrations. The concentration the midW mutant relaxation rates follows the same trend as the wild-type relaxation profile. A fast dependence of the midW mutant relaxation rates follows the same trend as the wild-type relaxation concentration-dependent relaxation rate and a slow concentration-independent relaxation rate were profile. A fast concentration-dependent relaxation rate and a slow concentration-independent also observed for wt-DHFR [14]. relaxation rate were also observed for wt-DHFR [14].

BC1000 5500 Slow Relaxation Rate Fast Relaxation Rate

) 900 ) 5000 -1 -1

800 4500 r = 0.882

on Rate (s Rate on 700 (s Rate on 4000

axati 600 r = -0 .020 axati 3500 Rel Rel

500 3000

80 100 120 80 100 120 Sum of Free Concentrations Sum of Free Concentrations FigureFigure 6. (A6.) Representative(A) Representative temperature temperature jump (T-jump) jump Trp(T-j fluorescenceump) Trp transientsfluorescence of midW47 transientsFolate of · withmidW47·Folate varying concentrations with varying of folate.concentrations The black of traces folate are. The double black exponential traces are double fits. (B exponential) Correlation fits. plot (B) forCorrelation the slow relaxation plot for the rates slow of relaxation midW47 Folate rates of versus midW47·Folate the sum of versus the concentration the sum of the of freeconcentration enzyme · andof free ligand.enzyme The and slow free relaxationligand. The rate slow is concentrationrelaxation rate independent, is concentration as shown independent, by the low as shown r value. by (Cthe) Correlation low r value. plot (C for) Correlation the fast relaxation plot for rates the fast of midW47 relaxationFolate rates versus of midW47·Folate the sum of free versus concentrations. the sum of · Thefree r valueconcentrations. close to one The confirms r value that close the fast to relaxationone confir ratems isthat strongly the fast concentration relaxation dependent.rate is strongly concentration dependent.

Assuming a simple two-state model where a ligand binds to an enzyme to form an enzyme– ligand complex, a linear fit of the fast relaxation rate as a function of the sum of free concentrations (Figures 6C and S4 B,D,F,H) provides additional information. The slope of the line is equivalent to kon and the y-intercept is equivalent to koff [23]. The kon and koff values for wt-DHFR and the midW mutants are reported in Table 3. The kon and koff are modulated in the midW mutants compared to wt-DHFR. While kon is similar to what has been reported in the literature for the DHFR–folate complex (57 ± 5 μM−1s−1) [24], the value of koff is two orders of magnitude larger (35 ± 12 s−1) [24]. Furthermore, comparing the Kd values determined by ITC experiments and the Kd values calculated from koff and kon reveals a discrepancy. The Kd values calculated from the T-jump experiments are approximately an order of magnitude larger than the Kd values determined from ITC. Taken together,

Molecules 2020, 25, 3819 8 of 13

Assuming a simple two-state model where a ligand binds to an enzyme to form an enzyme– ligand complex, a linear fit of the fast relaxation rate as a function of the sum of free concentrations (Figure6C and Figure S4B,D,F,H) provides additional information. The slope of the line is equivalent to kon and the y-intercept is equivalent to koff [23]. The kon and koff values for wt-DHFR and the midW mutants are reported in Table3. The k on and koff are modulated in the midW mutants compared to wt-DHFR. While kon is similar to what has been reported in the literature for the DHFR–folate complex (57 5 µM 1s 1)[24], the value of k is two orders of magnitude larger (35 12 s 1)[24]. Furthermore, ± − − off ± − comparing the Kd values determined by ITC experiments and the Kd values calculated from koff and kon reveals a discrepancy. The Kd values calculated from the T-jump experiments are approximately an order of magnitude larger than the Kd values determined from ITC. Taken together, these results indicate that the observed fast relaxation rate is not only due to ligand association/dissociation but is also convolved with associated conformation changes, such as the Met20 or FG loop dynamics. Loop motions within DHFR are critical for enzyme activity. In particular, two mutations have been studied that impact loop flexibility and consequently impact enzyme catalysis. When N23 at the end of the Met20 loop is replaced with a double proline sequence, millisecond fluctuations of the Met20 loop are abrogated [12]. When G121 on the FG loop is replaced with a valine, the pico-nanosecond and micro-millisecond motions of both the FG and Met20 loops are decreased [11]. Both of these studies used NMR relaxation experiments and thus were limited to time regimes accessible by NMR. Since T-jump spectroscopy is capable of measuring dynamics on the nanosecond to millisecond timescale, future studies with the N23PP or G121V mutants could use T-jump spectroscopy to fill in the time gap.

Table 3. The kon, koﬀ, and Kd for the midW mutants determined by T-jump transient data analysis compared to the Kd determined by isothermal titration calorimetry (ITC).

1 1 1 Enzyme kon (µM− s− ) koﬀ (s− ) Kd (µM) from T-jump Kd (µM) from ITC WTa 31 2 1100 200 35.5 6.8 4.7 ± ± ± midW22 38 20 3300 1550 88 61 3.7 ± ± ± midW30 51 6 3400 450 66 12 4.5 ± ± ± midW47 22 2 2000 240 92 14 12 ± ± ± midW74 53 3 2900 310 55.4 6.6 8.4 ± ± ± midW133 51 6 3100 550 60 13 9.5 ± ± ± a Reported values from Reddish et al. [14].

Although the T-jump transients of all five mutants show similar behavior with respect to concentration dependence, the relaxation rate of the slow phase varies based on the position of the Trp. The average relaxation rates of the slow phase are reported in Table4. Since the slow relaxation rate is concentration independent, the average rate is the combined average of the five replicates at the five ligand concentrations (n = 25). W47 and W74 are spatially close together; they are located at the top of DHFR when orientated as shown in Figure1. MidW47 and midW74 have close average 1 slow relaxation rates of approximately 730 s− . The slow phase relaxation rates for MidW22 and 1 midW30 are substantially faster, in excess of 1000 s− and are within one standard deviation of one another. Similarly, W22 and W30 are close together spatially; they are located near the substrate binding site. Lastly, W133 is located on a beta sheet near the distal FG loop, and midW133 has the 1 slowest average slow phase relaxation rate of 370 s− . To compare, the binary complex of wt-DHFR 1 with folate has a slow concentration-independent relaxation rate of 400 s− . This is slower than all but one of the concentration-independent rates observed in the midW mutants. A linear combination of the relaxation rates of midW mutants would not result in the relaxation rate observed in the wild-type enzyme unless the wt-DHFR transients were almost completely dominated by the W133 signal with very little contribution from the other four Trp residues. Our equilibrium fluorescence results show that midW133 has a low fluorescence intensity compared to the other midW mutants. Additionally, Ohmae et al. showed that when W133 is replaced with phenylalanine, the tryptophan fluorescence of DHFR is reduced by less than 15% [20]. Therefore, it is unlikely that W133 is dominating the Molecules 2020, 25, 3819 9 of 13

fluorescence signal in wt-DHFR. Instead, Trp–Trp interactions as well as the local environment of the Trp residues contribute to the wt-DHFR transients, similar to what we observed with the equilibrium fluorescence. Thus, the motions reported in the multi-Trp system are fundamentally different than the motions reported in the single Trp systems. We conclude that the wt-DHFR signal is dominated by Trp coupling effects due to relative changes in the distance and orientation of the Trp residues, which means that the wt-DHFR transients are reporting on global motions throughout the enzyme. Conversely, the signals from the midW mutants report on local motions that affect the environment of the single tryptophan in the enzyme, as evidenced by the dependence of the slow relaxation rate on the Trp location.

Table 4. Average relaxation rates of the slow phase for the binary folate complexes of all ﬁve midW mutants and wt-DHFR.

Enzyme midW22 midW30 midW47 midW74 midW133 wt-DHFR a Average Slow Rate (s 1) 2700 900 1800 960 740 90 720 180 370 90 400 70 − ± ± ± ± ± ± a Reported values from Reddish et al. [14].

By comparing the relaxation rates of the midW mutants, we can gain information about the environments of the individual Trp residues. Of particular interest are midW22 and midW30 because they both exhibit faster relaxation rates compared to the other midW mutants. The faster relaxation rate indicates that W22 and W30 are sensitive to conformational changes of a local ﬂexible structure. MidW30 has a lower melting temperature compared to wt-DHFR and the other midW mutants. During the T-jump experiments, samples are transiently heated to 36 ◦C, which is 6 ◦C lower than the melting temperature of midW30. It is possible that unfolding events may be playing a role. Another possibility is that W22 and W30 could be reporting on the same motion. W30 is located near the active site, and W22 is located at the end of the Met20 loop, which is known to change conformations during the catalytic cycle: opening and closing over the active site as well as protruding into the active site. The conformation of the Met20 loop can be modulated by ligand binding. In the binary complex with folate, the Met20 loop is in the occluded conformation, and in the apoenzyme, the Met20 loop is in the open conformation. Since W22 and W30 are on or near the Met20 loop, it is reasonable that they could be sensitive to the Met20 loop motions. The slow relaxation rates of midW22 and midW30 are approaching that of the fast concentration-dependent relaxation rate. The strong concentration dependence of the faster relaxation rate indicates that it is strongly coupled to ligand association/dissociation. The slow relaxation rate for midW22 and midW30 is not concentration dependent, which means that the slow relaxation rate is not coupled to ligand binding. This suggests that the observed slow relaxation rate is not the open to occluded transition. However, the slow relaxation event sensed by W22 and W30 is still likely related to Met20 loop ﬂuctuations due to the location of the Trp probes and the observed timescale. Loop motions typically occur in tens of nanoseconds to hundreds of microseconds, [1,25] and the slow relaxation events of midW22 and midW30 occur at approximately 300–500 microseconds. This is in contrast to what is observed with wt-DHFR and the other midW mutants where the slow relaxation events occur on the order of milliseconds.

3. Materials and Methods

3.1. Mutant DHFR Plasmid Generation, Protein Expression, and Purification C-terminal six histidine-tagged DHFR midW mutant plasmids were generated via the Custom Cloning Division within the Emory Integrated Genomics Core. In each midW mutant, all but one of the native tryptophans were mutated to phenylalanine (Phe). The prefix “mid” is old English for with; thus, midW22 is the DHFR mutant containing W22 with all other tryptophans mutated to phenylalanine. The midW mutants were expressed and purified as described previously for wild-type DHFR [14]. Briefly, BL21 (DE3) E. coli competent cells were transformed with the midW mutant plasmids, grown Molecules 2020, 25, 3819 10 of 13 with 100 µg/mL ampicillin, induced with 1 mM isopropyl β-d-thiogalactopyranoside (IPTG) overnight, and harvested by centrifugation. The pelleted cells were lysed by sonication, and the insoluble debris was removed by centrifugation. The supernatant was purified by loading on a HisPrep affinity column and washed to remove protein contaminants as well as endogenous NADPH. Finally, the midW protein was eluted with 500 mM imidazole, buffer exchanged, concentrated, and stored at 80 C. Protein − ◦ concentration was determined by absorbance at 280 nm with a predicted molar extinction coefficient of 1 1 10,810 M− cm− [26].

3.2. Relative Activity Assay The relative activities of the midW mutants compared to wild-type DHFR were measured as previously described [27]. Brieﬂy, 20 µL of 2.5 mM DHF were added to 980 µL of pre-equilibrated 10 nM DHFR and 51 µM NADPH in phosphate buﬀer (50 mM sodium phosphate, 100 mM NaCl, 5 mM 2-mercaptoethanol, pH 7). The rate of reaction was measured by monitoring the disappearance of the 340 nm absorbance band as NADPH was oxidized. The initial rates are reported in Table S1 as a percentage of the initial rate of wild-type DHFR.

3.3. CD Spectroscopy and Thermal Melts CD spectroscopy measurements were taken on a JASCO J-810 spectropolarimeter equipped with a PFD-425S Jasco temperature controller (Jasco, Inc., Easton, MD, USA) in a quartz cuvette with a 1 mm path length. Samples of wt-DHFR and midW mutants (15–30 µM) were prepared in 10 mM phosphate buﬀer, pH 7. The melting temperatures of the midW mutants were determined by monitoring the loss of ellipticity at 222 nm with increasing temperature. Measurements were taken in increments of 5 ◦C from 10 to 90 ◦C. To calculate the melting temperatures, the change in ellipticity at 222 nm versus temperature was ﬁt to a sigmoid.

3.4. Equilibrium Fluorescence Equilibrium fluorescence measurements of wt-DHFR apoenzyme and the midW mutants with their corresponding complexes (apoenzyme, E Folate, E NADP+, and E NADP+ Folate) were taken · · · · on a Horiba (Kyoto, Japan) Dual-Fl spectrofluorometer. Spectra were obtained by exciting with 280 nm light, corresponding to tryptophan absorbance. Due to the saturation limits of the fluorometer detector, wt-DHFR spectra were collected with an integration time of 0.5 s. The spectra of the midW mutants were collected with an integration time of 1.5 s. Each sample contained 3 µM of enzyme in phosphate buffer (50 mM sodium phosphate, 100 mM NaCl, pH 7). Samples included 6 µM of the appropriate ligands for the two binary complexes. For the tertiary complex, both 6 µM folate and 6 µM NADP+ were present. Spectra were collected in triplicate and were averaged together. To compare changes in fluorescence intensities between complexes, the entirety of the Trp fluorescence peaks were integrated. The integrated intensities of the enzyme–ligand complexes of each mutant are reported as a percentage of the corresponding apoenzyme fluorescence intensity. Additionally, temperature-dependent fluorescence spectra were collected from 12 to 45 ◦C in increments of 3 ◦C. The Trp fluorescence was integrated between 327 and 353 nm, which corresponds to the bandpass filter used in the temperature jump experiments. The resulting intensities were normalized to 1 with respect to the lowest temperature measured, and the intrinsic temperature dependence of tryptophan was subtracted as described in the supplemental information. These results are presented as change in fluorescence intensity.

3.5. Temperature Jump Fluorescence Spectroscopy Temperature jump fluorescence transients were collected and analyzed as described previously [14,28]. Briefly, the temperature jump was induced via a 2.09 µm, approximately 10 ns laser pulse at 12.5 Hz. The changes in Trp fluorescence were probed by exciting the sample with 280 nm light, focusing the emission through a bandpass filter (327–353 nm), and measuring the intensity with a photomultiplier Molecules 2020, 25, 3819 11 of 13

tube. The temperature jumps reported here are from 29 to 36 ◦C; this range is below the melting temperatures of the midW mutants, which range from 42 to 49 ◦C (Table1). The phosphate bu ffer that was used for the equilibrium fluorescence experiments was D2O exchanged for T-jump use. For each midW mutant, five samples were prepared containing 100 µM enzyme with varying concentrations of folate, up to 200 µM. The T-jump transience was collected in replicates of five. The intrinsic temperature-dependence of tryptophan was subtracted from each transient using approximately 200 µM free tryptophan as a reference. Then, the corrected fluorescence transients were normalized to 100 by dividing the entire transient by its initial fluorescence intensity, effectively reporting a percent change in fluorescence intensity over time. Finally, each transient was fit to a double exponential.

4. Conclusions In this study, we characterized the equilibrium Trp fluorescence and T-jump dynamics of five DHFR mutants, each with a single tryptophan residue. The equilibrium fluorescence confirms that each tryptophan residue is in a unique environment and that there are local quenching interactions. The wt-DHFR Trp fluorescence is not a simple combination of the individual tryptophans0 fluorescence due to Trp–Trp interactions such as excimer formation with W47 and W74. The T-jump dynamics demonstrate that all of the tryptophan residues are sensitive to ligand binding and dissociation in combination with associated conformational changes. Furthermore, the flexibility of DHFR is highlighted by the presence of slow microsecond–millisecond conformational changes throughout the enzyme that are not coupled to ligand association/dissociation processes relating to progress along the catalytic cycle. These fluctuations do not originate from a single collective motion, but instead arise from local motions with slightly different relaxation rates. Some of the relaxation rates are quite slow and likely do not impact catalysis. However, some relaxation rates, such as those observed in midW22 and midW30, are fast compared to the steady-state turnover rate of DHFR, and thus, they could be involved in conformational sampling and the search for reactive conformations. More generally, this study explores the utility of fluorescence spectroscopy applied to multi and single Trp systems. Trp fluorescence is often used as a probe for protein and enzyme studies. Trp is a naturally occurring amino acid, so endogenous Trp residues can be used as a probe or Trp can be easily incorporated into the protein of interest. Trp fluorescence is an ideal reporter because it is sensitive to the environment of the Trp. However, interpreting changes in Trp fluorescence can be challenging, because Trp is sensitive to local changes in its environment as well as various quenching interactions, Trp–Trp interactions, and Förster resonance energy transfer (FRET) to other fluorophores. Here, we demonstrated that multi-Trp systems can be used to follow global conformational changes as a result of Trp–Trp interactions. Furthermore, we showed that working with the corresponding single Trp mutants not only provides insight about the local conformational changes and environments of the individual tryptophans but also provides the necessary information to confirm that the multi-Trp system was reporting on global conformational changes. This type of coupled multi- and single fluorophore approach is broadly applicable and could be used to study the dynamics of other protein and enzyme systems.

Supplementary Materials: The following are available online at http://www.mdpi.com/1420-3049/25/17/3819/s1, Additional information on the activity and melting temperatures of the midW mutants, as well as the ITC results and temperature dependence of the slow and fast relaxation rates are available online. Author Contributions: Conceptualization, M.B.V. and R.B.D.; methodology, M.B.V.; validation, M.B.V., C.B., Q.L., and A.R.; formal analysis, M.B.V., C.B., and A.R.; investigation, M.B.V., C.B., Q.L., and A.R.; resources, R.B.D.; writing—original draft preparation, M.B.V.; writing—review and editing, M.B.V. and R.B.D.; visualization, M.B.V.; supervision, M.B.V. and R.B.D.; project administration, M.B.V. and R.B.D.; funding acquisition, R.B.D. All authors have read and agreed to the published version of the manuscript. Funding: This work was supported by National Institutes of Health Grant GM53640 (R.B.D.). Molecules 2020, 25, 3819 12 of 13

Acknowledgments: This study was supported in part by the Emory Integrated Genomics Core (EIGC), which is subsidized by the Emory University School of Medicine and is one of the Emory Integrated Core Facilities. Additional support was provided by the Georgia Clinical & Translational Science Alliance of the National Institutes of Health under Award Number UL1TR002378. The content is solely the responsibility of the authors and does not necessarily reflect the official views of the National Institutes of Health. Conflicts of Interest: The authors declare no conflict of interest.

References

1. Henzler-Wildman, K.; Kern, D. Dynamic personalities of proteins. Nature 2007, 450, 964–972. [CrossRef] 2. Mhashal, A.R.; Vardi-Kilshtain, A.; Kohen, A.; Major, D.T. The role of the Met(20) loop in the hydride transfer in Escherichia coli dihydrofolate reductase. J. Biol. Chem. 2017, 292, 14229–14239. [CrossRef] 3. Huang, Q.; Rodgers, J.M.; Hemley, R.J.; Ichiye, T. Extreme biophysics: Enzymes under pressure. J. Comput Chem 2017, 38, 1174–1182. [CrossRef] 4. Wang, Z.; Antoniou, D.; Schwartz, S.D.; Schramm, V.L. Hydride Transfer in DHFR by Transition Path Sampling, Kinetic Isotope Effects, and Heavy Enzyme Studies. Biochemistry 2016, 55, 157–166. [CrossRef] 5. Oyen, D.; Fenwick, R.B.; Stanfield, R.L.; Dyson, H.J.; Wright, P.E. Cofactor-Mediated Conformational Dynamics Promote Product Release From Escherichia coli Dihydrofolate Reductase via an Allosteric Pathway. J. Am. Chem Soc. 2015, 137, 9459–9468. [CrossRef][PubMed] 6. Singh, P.; Francis, K.; Kohen, A. Network of remote and local protein dynamics in dihydrofolate reductase catalysis. ACS Catal 2015, 5, 3067–3073. [CrossRef] 7. Hughes, R.L.; Johnson, L.A.; Behiry, E.M.; Loveridge, E.J.; Allemann, R.K. A Rapid Analysis of Variations in Conformational Behavior during Dihydrofolate Reductase Catalysis. Biochemistry 2017, 56, 2126–2133. [CrossRef] 8. Mhashal, A.R.; Pshetitsky, Y.; Eitan, R.; Cheatum, C.M.; Kohen, A.; Major, D.T. Effect of Asp122 Mutation on the Hydride Transfer in E. coli DHFR Demonstrates the Goldilocks of Enzyme Flexibility. J. Phys. Chem B 2018, 122, 8006–8017. [CrossRef] 9. Huang, Q.; Rodgers, J.M.; Hemley, R.J.; Ichiye, T. Adaptations for Pressure and Temperature Effects on Loop Motion in Escherichia coli and Moritella profunda Dihydrofolate Reductase. High. Press Res. 2019, 39, 225–237. [CrossRef][PubMed] 10. Sawaya, M.R.; Kraut, J. Loop and subdomain movements in the mechanism of Escherichia coli dihydrofolate reductase: Crystallographic evidence. Biochemistry 1997, 36, 586–603. [CrossRef][PubMed] 11. Boehr, D.D.; Schnell, J.R.; McElheny, D.; Bae, S.H.; Duggan, B.M.; Benkovic, S.J.; Dyson, H.J.; Wright, P.E. A distal mutation perturbs dynamic amino acid networks in dihydrofolate reductase. Biochemistry 2013, 52, 4605–4619. [CrossRef][PubMed] 12. Bhabha, G.; Lee, J.; Ekiert, D.C.; Gam, J.; Wilson, I.A.; Dyson, H.J.; Benkovic, S.J.; Wright, P.E. A dynamic knockout reveals that conformational fluctuations influence the chemical step of enzyme catalysis. Science 2011, 332, 234–238. [CrossRef][PubMed] 13. Ramanathan, A.; Agarwal, P.K. Evolutionarily conserved linkage between enzyme fold, flexibility, and catalysis. PLoS Biol. 2011, 9, e1001193. [CrossRef] 14. Reddish, M.J.; Vaughn, M.B.; Fu, R.; Dyer, R.B. Ligand-Dependent Conformational Dynamics of Dihydrofolate Reductase. Biochemistry 2016, 55, 1485–1493. [CrossRef][PubMed] 15. Roder, H.; Maki, K.; Cheng, H. Early events in protein folding explored by rapid mixing methods. Chem. Rev. 2006, 106, 1836–1861. [CrossRef][PubMed] 16. Rovo, P.; Linser, R. Microsecond Timescale Protein Dynamics: A Combined Solid-State NMR Approach. Chemphyschem 2018, 19, 34–39. [CrossRef] 17. Marcsisin, S.R.; Engen, J.R. Hydrogen exchange mass spectrometry: What is it and what can it tell us? Anal. Bioanal. Chem. 2010, 397, 967–972. [CrossRef] 18. Gotz, A.W.; Williamson, M.J.; Xu, D.; Poole, D.; Le Grand, S.; Walker, R.C. Routine Microsecond Molecular Dynamics Simulations with AMBER on GPUs. 1. Generalized Born. J. Chem. Theory Comput. 2012, 8, 1542–1555. [CrossRef] 19. Agarwal, P.K.; Billeter, S.R.; Rajagopalan, P.T.; Benkovic, S.J.; Hammes-Schiffer, S. Network of coupled promoting motions in enzyme catalysis. Proc. Natl. Acad. Sci. USA 2002, 99, 2794–2799. [CrossRef] Molecules 2020, 25, 3819 13 of 13

20. Ohmae, E.; Sasaki, Y.; Gekko, K. Effects of five-tryptophan mutations on structure, stability and function of Escherichia coli dihydrofolate reductase. J. Biochem. 2001, 130, 439–447. [CrossRef] 21. Gekko, K.; Yamagami, K.; Kunori, Y.; Ichihara, S.; Kodama, M.; Iwakura, M. Effects of point mutation in a flexible loop on the stability and enzymatic function of Escherichia coli dihydrofolate reductase. J. Biochem. 1993, 113, 74–80. [CrossRef][PubMed] 22. Kobayashi, K. A Handbook of Applied Statistics in Pharmacology; Correlation Analysis; CRC Press/Taylor Francis Group: Boca Raton, FL, USA, 2012. 23. Bernasconi, C.F. Relaxation Kinetics; Academic Press: New York, NY, USA, 1976. 24. Posner, B.A.; Li, L.; Bethell, R.; Tsuji, T.; Benkovic, S.J. Engineering specificity for folate into dihydrofolate reductase from Escherichia coli. Biochemistry 1996, 35, 1653–1663. [CrossRef][PubMed] 25. Schramm, V.L. Transition States and transition state analogue interactions with enzymes. Acc. Chem. Res. 2015, 48, 1032–1039. [CrossRef] 26. Gill, S.C.; von Hippel, P.H. Calculation of protein extinction coefficients from amino acid sequence data. Anal. Biochem. 1989, 182, 319–326. [CrossRef] 27. Chen, S.; Fahmi, N.E.; Wang, L.; Bhattacharya, C.; Benkovic, S.J.; Hecht, S.M. Detection of dihydrofolate reductase conformational change by FRET using two fluorescent amino acids. J. Am. Chem Soc. 2013, 135, 12924–12927. [CrossRef][PubMed] 28. Davis, C.M.; Reddish, M.J.; Dyer, R.B. Dual time-resolved temperature-jump fluorescence and infrared spectroscopy for the study of fast protein dynamics. Spectrochim. Acta A Mol. Biomol. Spectrosc. 2017, 178, 185–191. [CrossRef][PubMed]

Sample Availability: Samples of the compounds are not available from the authors.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).