import numpy as np

p=998244353

def mul(a, b, MOD=p):
    fft=np.fft
    def submul(f,g):
        L=len(f)+len(g)-1
        return fft.irfft(fft.rfft(f,L)*fft.rfft(g,L),L)%MOD
    al=a&(1<<15)-1
    bl=b&(1<<15)-1
    ah=a>>15
    bh=b>>15
    x,y,z=submul(al,bl),submul(ah,bh),submul(al+ah,bl+bh)
    x,y,z=map(lambda v:(v+.5).astype(np.int64),[x,y,z])
    return (x+(y<<30)+((z-x-y)<<15))%MOD

fac=[1, 1]
inv=[1, 1]
ifac=[1, 1]
for i in range(2, int(2e5)):
    fac.append(fac[-1]*i%p)
    inv.append(p-inv[p%i]*(p//i)%p)
    ifac.append(ifac[-1]*inv[i]%p)

def fpow(a, n):
    if n==0:
        return np.array([1])
    else:
        ret=fpow(mul(a, a)[:10**5+10], n//2)[:10**5+10]
        if n%2==1:
            ret=mul(a, ret)[:10**5+10]
        return ret

def comb(n, k):
    return fac[n]*ifac[n-k]%p*ifac[k]%p

N,M,K=map(int,input().split())

f=np.array([0]*(10**5+10))
for i in range(len(f)):
    f[i]=ifac[i]
f[0]=0
f=fpow(f,K)

ans=0

for i in range(K,N+1):
    ans+=comb(N, i)*comb(M, K)%p*pow(M,N-i,p)%p*f[i]%p*fac[i]%p
    ans%=p
print(ans)